The global stock market in the spring of 2026 is once again revolving around semiconductors. The tickers on the New York Stock Exchange and Nasdaq are consistently glowing red, with investors focusing on a single term: artificial intelligence (AI).
The recent surge in semiconductor stocks on the U.S. market is not merely a reaction to positive news; it signals that the new industrial revolution driven by AI is shaking financial markets, industrial structures, national strategies, and geopolitics simultaneously.
Micron's stock skyrocketed by 15% in a single day, while AMD and Intel saw increases of nearly 10-15%. The market is beginning to view semiconductors as the most critical strategic asset since oil.
Particularly, the atmosphere in the memory semiconductor market has changed dramatically. Once considered a 'chicken game industry,' where companies like Samsung Electronics and SK Hynix would increase supply, causing prices to plummet, the memory sector has transformed. D-RAM prices collapsed during global economic turmoil, leading to extreme fluctuations in semiconductor companies' operating profits.
However, with the advent of the AI era, the market is starting to view the memory industry from an entirely different perspective.
Today, what AI data centers require is not just simple storage but ultra-fast memory capable of processing trillions of operations per second. The key component is High Bandwidth Memory (HBM), which acts as the brain of AI servers when combined with NVIDIA GPUs. If the GPU is the engine, HBM serves as the blood vessels supplying oxygen.
As AI models grow larger, the demand for HBM is skyrocketing. Companies like OpenAI, Google, Meta, Amazon, xAI, and major Chinese tech firms are all competing in the AI data center race, leading to a near depletion of global HBM supply.
SK Hynix has emerged as a key player in NVIDIA's supply chain, earning recognition as one of the biggest beneficiaries of the AI era due to its strong early-mover advantage in the HBM market. Samsung Electronics is also making astronomical investments to secure next-generation HBM4 and advanced packaging capabilities.
The global semiconductor industry is shifting from mere corporate competition to a national all-out war.
The U.S. is expanding its advanced semiconductor manufacturing base through the CHIPS Act, while China is striving for semiconductor independence as a matter of national survival. Japan is collaborating with TSMC on a semiconductor revival project centered in Kyushu, and countries like India and those in the Middle East are entering the competition to attract AI data centers and advanced chips.
Semiconductors are no longer just electronic components; they are military power, financial systems, and core assets of national security.
In fact, most modern weapons operate on semiconductors. Drones, missiles, satellites, AI reconnaissance systems, autonomous weapons, and cyber warfare systems all require advanced chips to function. The U.S. has tightened export controls on advanced GPUs to China for this very reason, as AI semiconductors are at the heart of future hegemony. Additionally, the structural characteristics of the AI industry are further stimulating the semiconductor supercycle.
In the past, during the internet era, even as the number of servers increased, the growth rate slowed beyond a certain point. However, AI is different. As AI models evolve, they require more data and computational power. The number of GPUs needed to train a single GPT-scale large language model exceeds that of past supercomputers.
Consequently, Wall Street now views NVIDIA not just as a semiconductor company but as the “oil supplier of the AI era.” GPUs have become the new crude oil, and HBM serves as the pipeline transporting that oil. The challenge is that this transformation is dramatically stimulating the market's imagination.
Recently, the mood on Wall Street resembles the gold rush of the 19th century. There is a prevailing sentiment that “any AI-related company will rise.” Companies involved in AI server equipment, power supply, and cooling solutions are all experiencing surges. The market for liquid cooling systems and power infrastructure for data centers is also expanding rapidly.
In Texas and Arizona, as well as in Saudi Arabia and the UAE, massive AI data center construction projects are being announced. Some projects are projected to require power equivalent to that of a nuclear power plant. This reveals that AI is not merely a software industry but a massive energy-consuming industry.
Notably, the capital expenditures (CAPEX) of major U.S. tech firms have reached historic levels. Annual AI-related investments by Microsoft, Meta, Amazon, and Google exceed the budgets of many countries. These companies are entering long-term supply contracts and pre-order agreements to avoid falling behind in the semiconductor procurement race.
Micron's recent surge is also a reflection of this trend. The market is not merely focused on “good performance” but is paying attention to the potential for a structural long-term memory shortage. Unlike past cycles that saw a brief rise followed by a downturn, there is a growing belief that the expansion of AI infrastructure could sustain growth for several years.
However, the market's enthusiasm also signals potential risks. A significant portion of the S&P 500's rise is concentrated in a small number of tech stocks, which is crucial to note. The market is not rising healthily; rather, a few AI stocks are driving the index higher. This suggests that the strength of the rally may be weaker than anticipated.
Particularly, the recent surge in call option purchases by retail investors is reminiscent of patterns seen during past bubbles. While the belief that AI will transform human civilization may be accurate, financial markets have historically tended to overprice future expectations.
During the late 1990s internet bubble, the assertion that “the internet will change the world” was true. Indeed, the internet did transform the world. However, many internet company stocks soared to unrealistic heights, ultimately leading to a bubble burst.
Today's AI market carries similar risks. Energy issues are a key variable. AI data centers consume enormous amounts of power. This is why countries worldwide are discussing the reactivation of nuclear power plants and the expansion of power grids. As the AI industry grows, issues related to electricity, gas, and oil will become increasingly significant.
This is why Wall Street is currently paying attention to the Strait of Hormuz. If geopolitical instability in the Middle East leads to a surge in oil prices, the cost structure of the AI industry could be disrupted. While the market is currently buoyed by inventories and strategic reserves, a long-term instability in energy supply chains could also impact the AI rally. Ultimately, the global semiconductor market stands at the intersection of technology, finance, geopolitics, energy, and military strategy during a monumental civilizational transition.
In this context, investment sages like Sir John Templeton would caution against the crowd's enthusiasm. He famously stated, “The greatest returns come at the most pessimistic moments.” Conversely, when everyone is caught up in optimism, he advised skepticism. Templeton warned that the phrase “this time is different” is often used when humans are at their most vulnerable.
Currently, Wall Street is effectively shouting “this time is different” regarding AI. If Templeton were here, he might acknowledge the long-term potential of the AI revolution but would likely view the market's excessive greed and crowd psychology with a critical eye.
Warren Buffett would pose a slightly different question: “Will that company still be making money in ten years?” Buffett focuses on cash flow, market dominance, and management's capital allocation abilities rather than technology itself. This is why he has been cautious about tech stock investments but has made long-term investments in Apple, seeing it as more than just a tech company but as a powerful consumer ecosystem with brand dominance.
Buffett also famously remarked, “When the tide goes out, you can see who has been swimming naked.” In times of abundant liquidity and strong AI enthusiasm, anyone can appear brilliant. However, when interest rates fluctuate, energy prices soar, and economic slowdowns occur, the differences between truly competitive companies and those that are merely bubbles become apparent.
Historically, all technological revolutions have followed similar patterns. The railroad revolution, the automobile revolution, and the internet revolution all saw massive investment booms in their early stages, but only a handful of companies emerged as ultimate winners, while the rest vanished. Ultimately, the crucial question is not whether “AI will change the world,” but rather “who will survive through that change.”
Eastern classics offer remarkable insights here. Laozi stated in the Dao De Jing, “What is full will spill over; what is sharp will not last long.” He also warned, “An excessively sharp sword cannot be preserved for long.” Markets have always collapsed when they overflow. Unchecked human desires sow the seeds of their own destruction.
Conversely, the I Ching speaks of the principles of change: “When one reaches the limit, change occurs; change leads to communication; communication leads to longevity.” This means that when one reaches a dead end, change must occur, and only through change can a path be opened that endures.
Today, the global economy stands at the threshold of that change. AI has the potential to revolutionize human productivity. However, it also casts a long shadow of energy shortages, geopolitics, financial bubbles, and hegemonic competition. The AI hegemony war between the U.S. and China is not merely an industrial competition; it represents a new cold war encompassing semiconductors, energy, data, military power, and financial systems.
In this process, South Korea's position is far from insignificant. Samsung Electronics and SK Hynix are at the core of the global AI supply chain. As the world races toward AI, the strategic value of South Korean semiconductors is likely to increase. However, South Korea must also remain vigilant. The moment the illusion of a semiconductor supercycle leads to an over-concentration of all industrial structures on AI and semiconductors, the risks will grow. Without securing diversity in the industrial ecosystem, energy security, and financial stability, South Korea will inevitably be shaken by external shocks.
Ultimately, the market is a mirror of human civilization. Greed and fear, innovation and illusion, technology and desire all intertwine. The current semiconductor rally is undoubtedly a signal of changing times. However, it is also a dangerous heat generated by human crowd psychology.
John Templeton would likely say, “Stay calm when the crowd is enthusiastic.” Warren Buffett might add, “Even good companies become bad investments if bought at too high a price.” And the ancient texts of the East quietly whisper, “Heaven and earth are not in a hurry, yet they accomplish everything.”
The era of AI and semiconductors is not over; rather, it is likely just beginning. However, the greater the era, the deeper the need for restraint and insight. True masters do not lose their center even amid enthusiasm. This is the harsh truth that Wall Street's history has repeatedly proven.
※ This article was generated using generative AI and has been reviewed by an editor.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.